Machine Learning Engineer, Performance Tooling
Core
Build self-service performance tools for the AI stack (models, compilers, runtimes, accelerators) to measure, predict, and optimize latency, throughput, and compute costs.
Role type
Senior IC machine learning engineer (performance tooling)
Builds
Scalable performance profiling and monitoring tools for model training and inference
Domain
AI infrastructure, deep learning, cloud and embedded hardware
Deliverable
production ML models | product features
Required skills
Python, PyTorch, performance engineering, profiling, roofline analysis, latency optimization, throughput optimization, data analysis, system design
Preferred skills
cross-stack AI architecture knowledge, quantitative communication
Technologies
PyTorch, Python
Responsibilities
Design and build reliable, self-service performance tools; Shape measurement and prediction standards for AI performance; Model theoretical peak vs. achieved performance; Predict latency, memory, and compute cost of changes; Own monitoring and regression alerting; Collaborate with training and runtime engineers to set performance targets